• Please review our updated Terms and Rules here

Drive Error 21 when trying to boot an 11/73

SteveJ

New Member
Joined
Oct 16, 2022
Messages
7
Hi all,

I've been a lurker for a while, but now I have a question...

I have a PDP-11/73 in a BA23 chassis (full details below), which I acquired about 40 years ago when an auction house was selling office stuff from a failed business. I got it for next to nothing since they had no idea what it was. It was running RT-11, I'd have preferred RSX-11M but although I had (still have, in fact) the RSX distribution on RL02 I never had a drive that could read them. I did get it running Ultrix 2.1, which it what it booted 25 years ago when last shut down. Around then I had bigger & faster Unix boxes, the PDP got put in a cupboard and neglected.

I remembered it when clearing out for a house move, and decided to see if it still worked. Wary of the likely effects of being idle for so long I replaced all the electrolytics in the PSU, and the notorious RIFA caps, before powering it up. Power seems good, correct voltages and looks clean on a 'scope, but although the system started fine, ran the self-tests, it won't boot from disk. Any attempt to access the disks, either the RD52 with Ultrix or bootable RT11 floppies in the RX50, produces familiar mechanical noises, but eventually just the message:
Drive error
Error 21

The RD52 isn't showing symptoms of a head crash, no metallic 'pinging', and the drive belt in the RX50 seems fine, everything moves OK. I did clean the RX50 heads, but I can't find any information online about "Error 21". Can anyone shed any light on this? It wouldn't surprise me if the RD52 were dead, or if the actual floppies had lost data, but it would be helpful to know for sure.

I'm tempted to try the floppy replacement device from "Gotek" with the flashfloppy software (will that work with an RQDX2?), or perhaps the tu58fs tape emulator if I can get an extra serial port for my PC. I'm a bit reluctant to spend too much more money on what might be a dead end but would that be worth a try? If so, what sort of disk/controller diagnostics could I try?

I do have a DELQA card as well, maybe I could boot it over the network? I'm very familiar with doing that on larger, newer, Unix systems but I don't know if the 11/73 can do that, or what setup I'd need to do on the host.

The system HW config is:
BA23, H9278 backplane. Installed cards (in this order) are:
  1. M8190-AB KDJ11-BB (quad card)
  2. M7551-AF MSV11-QA 5017547-01-C1-P2 4MB Memory (quad card)
  3. M8043 DLV11-J 4-port serial interface (double card, CD slots empty)
  4. Emulex TU0210401 TC02 tape controller (emulates TS11) (quad card)
  5. M8639-YB RQDX2 RDRX disk controller 50-15649-02. Firmware 189E5. (quad card)
is it still the case that the RQDX2 needs to be the last card in the chassis? If so I'd need to swap out the DLV11 for the DELQA since I don't have a bus grant card to put the DELQA in an AB/AB slot.

Any help diagnosing this would be much appreciated.

Thanks
Steve
 
Hi all,

I've been a lurker for a while, but now I have a question...

I have a PDP-11/73 in a BA23 chassis (full details below), which I acquired about 40 years ago when an auction house was selling office stuff from a failed business. I got it for next to nothing since they had no idea what it was. It was running RT-11, I'd have preferred RSX-11M but although I had (still have, in fact) the RSX distribution on RL02 I never had a drive that could read them. I did get it running Ultrix 2.1, which it what it booted 25 years ago when last shut down. Around then I had bigger & faster Unix boxes, the PDP got put in a cupboard and neglected.

I remembered it when clearing out for a house move, and decided to see if it still worked. Wary of the likely effects of being idle for so long I replaced all the electrolytics in the PSU, and the notorious RIFA caps, before powering it up. Power seems good, correct voltages and looks clean on a 'scope, but although the system started fine, ran the self-tests, it won't boot from disk. Any attempt to access the disks, either the RD52 with Ultrix or bootable RT11 floppies in the RX50, produces familiar mechanical noises, but eventually just the message:
Drive error
Error 21

The RD52 isn't showing symptoms of a head crash, no metallic 'pinging', and the drive belt in the RX50 seems fine, everything moves OK. I did clean the RX50 heads, but I can't find any information online about "Error 21". Can anyone shed any light on this? It wouldn't surprise me if the RD52 were dead, or if the actual floppies had lost data, but it would be helpful to know for sure.

I'm tempted to try the floppy replacement device from "Gotek" with the flashfloppy software (will that work with an RQDX2?), or perhaps the tu58fs tape emulator if I can get an extra serial port for my PC. I'm a bit reluctant to spend too much more money on what might be a dead end but would that be worth a try? If so, what sort of disk/controller diagnostics could I try?

I do have a DELQA card as well, maybe I could boot it over the network? I'm very familiar with doing that on larger, newer, Unix systems but I don't know if the 11/73 can do that, or what setup I'd need to do on the host.

The system HW config is:
BA23, H9278 backplane. Installed cards (in this order) are:
  1. M8190-AB KDJ11-BB (quad card)
  2. M7551-AF MSV11-QA 5017547-01-C1-P2 4MB Memory (quad card)
  3. M8043 DLV11-J 4-port serial interface (double card, CD slots empty)
  4. Emulex TU0210401 TC02 tape controller (emulates TS11) (quad card)
  5. M8639-YB RQDX2 RDRX disk controller 50-15649-02. Firmware 189E5. (quad card)
is it still the case that the RQDX2 needs to be the last card in the chassis? If so I'd need to swap out the DLV11 for the DELQA since I don't have a bus grant card to put the DELQA in an AB/AB slot.

Any help diagnosing this would be much appreciated.

Thanks
Steve
The RQDX2 is an etch fix to the RQDX1 to fix the "has to be the last card" botch. Note that some RQDX1 boards have PROMs from the 2, just to confuse you.

I'd check the ribbon cable from the RQDX to the backplane and the cables from the backplane to the drives. If your BA23 has a 6 button control panel, things can get interesting. Most shipped with the 4 button panel.

Check the power supply harness at both the backplane and PSU. IIRC, you want the one with individual multiple colored wires and white connectors. Ribbon power cable is a no-no, as is a cable with individual wires but with yellow connectors. If you have either of the last 2, look for signs of heat damage on the connectors.
 
Didn't know that! No other differences?
There's two RQDX1 board variants - M8639 and M8639-YA. The RQDX2 is M8639-YB. DEC used -Yx variants to denote minor variations in configuration / firmware / operational mode. On memory, generally in MSxxx-yz, xxs is for the processor family and optionally indicates parity / ECC, y is the board capacity (which can be a difference of 16 / 64 / 256Kbit chips or a partially vs. fully-stuffed board, and z is a code for the memory chip manufacturer.

We referred to the non-YA RQDX1 as the "mud board" because DEC used a pale brown solder mask.

RQDX2 EPROMs will run in the -YA variant of the RQDX1. IIRC, the non-YA variant of the RQDX1 was limited to 8KB EPROMs while the -YA could use 8KB or 16KB (the firmware developed a serious case of bloat).

Note that once newer EPROMs "talk" to a drive, that drive will generally no longer work with older EPROMs, even in the same controller board. There are a number of qualifications to that which are too byzantine to go into here - just assume that upgrading is a one-way operation.

The RQDX3 is a better board in just about every way - size, speed, user-configurable geometry, etc.
 
The RQDX2 is an etch fix to the RQDX1 to fix the "has to be the last card" botch. Note that some RQDX1 boards have PROMs from the 2, just to confuse you.

I'd check the ribbon cable from the RQDX to the backplane and the cables from the backplane to the drives. If your BA23 has a 6 button control panel, things can get interesting. Most shipped with the 4 button panel.

Check the power supply harness at both the backplane and PSU. IIRC, you want the one with individual multiple colored wires and white connectors. Ribbon power cable is a no-no, as is a cable with individual wires but with yellow connectors. If you have either of the last 2, look for signs of heat damage on the connectors.
Thanks Terri & others. Ribbon cables all test OK and are well-seated, and the PSU harness is the individual wires with white connectors, which show no signs of overheating. The control panel is the 4-button type.
It does seem like my next step is to get the system to boot via the tu58fs emulator, and see if XXDP+ can shed any light. In the meantime I'll pull the PSU again & double-check that I haven't broken a cable or caused any other issues.

Steve
 
I finally got back to this. I was able to get the system to boot via the tu58fs emulator, and get XXDP+ running, so basically it seems OK, at least as far as CPU/RAM goes. Running the XXDP+ RX50 and RD52 exercisers, set not to write to the media, just produced read errors from both RD52 and RX50. Trying to write to an RX50 diskette gave write errors. I'm reluctant to try & reformat or otherwise write to the RD52 since it should still have a bootable Ultrix image on it.

Next step was to get a Gotex floppy emulator, which I connected in place of the RX50. It seems to be recognised, when booting I can see it step through tracks, but at best I get "no bootable media" errors. I think I correctly created the images & configured the Gotek, but am not sure. I do see different behaviour with different images, which makes me think that the system is seeing the images but isn't finding boot info on them. That leaves me with a couple of possible issues: either I don't have valid bootable RX50 images, or there's a problem with the RQDX2 (there are no diagnostic lights lit after the self-test completes, though). I haven't yet tried XXDP+ with an image on the Gotex, that's next.

Would any kind soul be able to point me to a known good RX50 boot image, RT11 or similar, that is suitable for the Gotek?

Also, are there any obvious known weaknesses for the RQDX2? I've heard of people having to replace the driver chip for the RD5x, but since both RX and RD devices give errors I'd guess that any possible fault on the board is further upstream, where the two data paths converge.
 
Back
Top